Probabilistic Physics-Informed Graph Convolutional Network for Active Distribution System Voltage Prediction

Tong Su, Junbo Zhao, Yansong Pei, Fei Ding

Research output: Contribution to journalArticlepeer-review

3 Scopus Citations


This letter proposes a novel data-driven probabilistic physics-informed graph convolutional network (GCN) for active distribution system voltage prediction with PVs and EVs. It leverages both measurements and network topology to accurately and efficiently predict node voltages without the need for an accurate distribution system power flow model. The dropout-enabled Bayesian inference is developed to achieve uncertainty quantification of the voltage prediction. Thanks to the network model embedding, it also has robustness against topology changes, a key difference with existing machine learning-based approaches. Comparison results with other state-of-the-art machine learning methods on a realistic 759-node distribution system demonstrate that the proposed method can achieve better accuracy and robustness under different scenarios.
Original languageAmerican English
Pages (from-to)5969-5972
Number of pages4
JournalIEEE Transactions on Power Systems
Issue number6
StatePublished - 2023

NREL Publication Number

  • NREL/JA-5D00-87747


  • distribution system
  • physics-informed graph convolutional network (GCN)
  • probabilistic voltage prediction


Dive into the research topics of 'Probabilistic Physics-Informed Graph Convolutional Network for Active Distribution System Voltage Prediction'. Together they form a unique fingerprint.

Cite this